Duplicate secret causing pod restart loop
hlx-a1 opened this issue · comments
Your environment
Chart Version: 1.15.0
Helm Version: 3.15.4
Kubernetes Version: v1.29.4+k3s1
What happened?
The presence of a duplicate secret in a 1P vault caused a container that mounted one of the copies to be trapped in a restart loop every 3 minutes. This happened because every 3 minutes, the 1P connect operator polled the 1P server, falsely detecting an update, and scaling the app down and up from 0 replicas.
What did you expect to happen?
I would expect the container to not be restarted if none of the secrets it mounts have changed. Note: we have operator.1password.io/auto-restart=true per OnePasswordItem.
Steps to reproduce
- Create a OnePasswordItem with operator.1password.io/auto-restart=true
- Create a deployment that mounts it
- Duplicate the entry that the OnePasswordItem references
- The deployment should be scaled down to 0 and back up every polling interval